Snowflake brings order to unstructured data with AI
- June 28, 2023
- 0
Snowflake uses an LLM in its data cloud that can read unstructured data, which should enable more and faster insights. Snowflake itself puts intelligence into its data cloud.
Snowflake uses an LLM in its data cloud that can read unstructured data, which should enable more and faster insights. Snowflake itself puts intelligence into its data cloud.
Snowflake uses an LLM in its data cloud that can read unstructured data, which should enable more and faster insights.
Snowflake itself puts intelligence into its data cloud. The data specialist integrates his own Large Language Models (LLMs) in its platform. The company made the announcement during the annual Snowflake Summit in Las Vegas.
Specifically, Snowflake integrates AI through an LLM from Applica: a company it acquired in 2022. The model is able to read unstructured data such as data in PDFs of invoices. In addition, the AI can quickly recognize things like amounts or contract terms. This functionality is called Snowflake Document AI. During an on-stage demo, the company emphasizes that the model doesn’t need fine-tuning to work properly.
You can ask questions to the AI, which searches the unstructured data for an answer. “You can ask questions about your documents in Snowflake.
Christian Kleinerman, SVP Product Snowflake
Document AI works with a simple visual interface that understands natural language. You can ask questions to the AI, which searches the unstructured data for an answer. “You can ask questions about your documents in Snowflake,” says Christian Kleinerman, SVP for product. “In doing so, we are responding to our customers’ desire to derive more value from their unstructured data.”
The integration is likely to be the first in a series of future innovations centered around unstructured data. Snowflake predicts that within five years about 90 percent of the world’s data will be unstructured, and who will cater for that? New AI models will help analyze images, video and audio in the future.
“Some features like document analysis are very important,” explains Benoit Dageville, co-founder of Snowflake. “We will develop such capabilities ourselves on our platform because we can do it very efficiently. For other AI implementations, we give users the tools to build and integrate them themselves.”
Dageville compares the approach to apps on a phone: “Apple develops some apps, but most of the functionality comes from other developers.” It’s not yet clear which AI capabilities Snowflake plans to adopt after Document AI.
Document AI is available as a private preview. Snowflake also ensures that customers can also use their data to train new AI models. This is done through a partnership with Nvidia that was also recently announced. On the technical side, container integration plays an important role. Snowflake is also deepening its collaboration with Microsoft.
Source: IT Daily
As an experienced journalist and author, Mary has been reporting on the latest news and trends for over 5 years. With a passion for uncovering the stories behind the headlines, Mary has earned a reputation as a trusted voice in the world of journalism. Her writing style is insightful, engaging and thought-provoking, as she takes a deep dive into the most pressing issues of our time.